Home »

Dear Sir

Nearly all research in science today requires the management and calculation of uncertainty, and for this reason statistics-the science of uncertainty-has become a crucial partner for modern science. Statistics is serious business. Still, much humor can be found in statistics and the people who study it-statisticians. What follows is a letter to the editor from the October 1963 issue of The American Statistician that displays this lighter side of statistics. May you all enjoy a good laugh!

Dear Sir:


Two years ago you'd have said I was a normal, unsuspecting female, with never a higher-mathematical thought in my head beyond trying to balance the check book, or wondering how I could squeeze that standing rib roast out of what's left in the food money. Whenever I heard a "t"-table mentioned, I'd say "No, honey, it's a coffee table we need." If the matter of a "log" table came up, I agreed that it would look fine in our rec room, but I drew the line at that contingency table—any woman knows it simply wouldn't fit in with our Danish Modern!

Then came a change in my engineer-husband's job to Quality Control, and slowly, sneakily, almost without my realizing it, Statistics began to worm its confusing way into my otherwise happy life, which consisted mainly of the care and cleaning of our five products. (I realize now that that's what they call the result of a multiplier and a multiplicand.)

To be sure, it began innocently enough, when a fellow engineer came to the house to go over a few things with my husband. I wasn't really eavesdropping. After all, when a gal who's been married for eighteen years hears her husband talking about curves, her antennae automatically go out! At first I assumed they were discussing the new office secretary, but when they threw in names like Diocles' Cissoid, Pascal's Limacon, and that Witch of Agnessi's, I was shocked. They were actually mentioning the names of the "other women" some of the fellows were seeing!

They stopped gossipping soon, and began talking tests. I figured, "Aha, now I can get in on the conversation," because I'd given hundreds of Binets, Wechslers, Arthurs, and even dabbled in the Rorschach. I was so wrong—these tests I'd never even heard of! Take the Null Hypothesis of Homogeniety of the Kullback Leibler, for example; don't they sound like a new way of pasteurizing milk or a different method of artificial respiration?

When they started arguing about the sequence of the various steps in solving certain equations, all I could think of was "watch out for that first step—it's a beaut!" Then as our friend was leaving, he asked my husband which word he had unscrambled at the last Quality Control seminar to win the scroll, and when my husband said "Homoscedasticity," I interrupted with "Gesundheit! How about taking a hot bath, two aspirin, and getting to bed!"

After that, as we journeyed hand in hand along the statistical trail, I began typing copies of articles for him, deceptively simple at first. Certain things gradually became clear to me, like that an asymptotic distribution is what he likes when it comes to cake or pie. And that the ambition of every square is probably to become a Chi square and deviate as standardly as the best of them. Also, that statistics can be taken with tongue in cheek, as many laymen do, or with foot in mouth.

Sometimes the titles of the papers he asked me to type sounded almost lyrical, like poetry— "On the Interpretation of from Contingency Tables," or "On the Meaning of Precision and Accuracy." I often wonder what would happen if I slipped in the wrong title, like "On assuming a facade of intelligence while your husband explains statistics."

The actual content of these statistical articles can throw you, if you're not careful. Take this little gem, for example: "If we assume beforehand a 5% level of significance, then since the critical region for 11 degrees of freedom at a 5% level of significance is the region with the values of ?2 greater than 19.68, we see that the two sample frequency counts may very reasonably be taken to have come from the same population." Makes you want to say "SO they live in the same town, so what?!"

And the functions they've got! Bessel, gamma, hyperbolic, exponential, inverse trigonometric and plain trigonometric. Add to these words in the vernacular like commutation, correlation, integration, operation—it's got quite a beat, and if they just added a shoo—be—do or two, it might even sell!

Books on statistics are actually quite deceptive. Once you're "hooked" you start looking up what you want to know, but frankly, under "Proportion Formulas" they don't have a thing about ideal measurements, diets, or reducing plans. Sometimes, too, they sound like a couple of women gossipping over the back fence— "the test is due to Pearson, but Fisher and Snedecor say-and then there's Herdan's opinion—etc."

The real curve they throw you at the end of these gems is where the author, backed up by fifty or sixty references, suggests that the reader go through the actual steps of the preceding calculations to see for himself the simplicity of the method—oh, those dreamers!

Speaking of references, usually there isn't a familiar name among them, until I chanced upon Good, I.J., and immediately I thought "I'll bet he's related to Chester on 'Gunsmoke'!" Ironically enough, this paper was entitled "Saddle-point methods for the multinomial distribution."

Statisticians themselves are quite serious about this whole business. A big wheel in Statistics once asked me if there really wasn't some particular number that appealed to me emotionally, that I felt especially comfortable with. I suppose I should have taken him seriously, but I answered "Only the numbers on my husband's pay check, but I imagine the number 8 would appeal to most men because it's so round, so firm, so fully packed." Funny, he hasn't spoken to me since!

Statistically, I guess I'm not even average. The average woman doesn't like boxing and baseball—I do. The average woman drives a car—I don't. But I can give you this important statistic about my husband: except for his interest in statistics, he is strictly a three sigma man, or, translated from statistical parlance, out of this world!

Sincerely yours,
Marian W. Quirk